[Insight-users] contrast enhancement filter

[Luis Ibanez]

Hi Orientation f.,


1) Contrast enahncement is only useful for segmentation if it is
    computed locally. Otherwise it is only misleading for the human
    operators.


2) The GeodesicActiveContourLevelSetImage filter in ITK is *already*
    implemented using a NarrowBand approach.


    Regards,



       Luis




---------------------
orientation f wrote:
> Hi,
> In order to gain a good segmentation result, I want  the region of 
> interest to be as distinguishable as possible in intensity from the 
> surrounding regions. Is there any image contrast enhancement filters can 
> meet this requirment?
> Another question: In level set method, the evolving contour updates over 
> all image domain, so the computing efficiency is low. The narrow band 
> method updates the contour within a given band width, which improves the 
> convergence speed greatly. But I don't know how to cooperate the narrow 
> band method with the common level set method, such as implement the 
> geodesic active contour with narrow band method?
> Thanks!